A Framework for Automated Extraction and Classification of Linear Networks

The authors conclude that the adoption of an object-oriented geospatial database has allowed complex discriminating characteristics of objects to be dynamically extracted, effectively enabling objects to classify themselves.

Abstract

Linear features, including roads, are important components of geographic information system (GIS) databases. This article outlines a framework for extracting networks of linear features from imagery using an object-oriented geodata model. The Automated Linear Feature Identification and Extraction (ALFIE) uses a control strategy to automate the process flow. The authors discuss the flexibility of the resulting system, which incorporates a toolkit of algorithms and imagery to extract linear features and utilizes contextual information to allow evidence of class membership to be built up from a variety of sources. The classification algorithm employs a Bayesian modeling approach that incorporates both geometric and photometric information. In this approach, five key discriminators are identified: width, width variation, sinuosity, spectral value, and spectral value variation. The authors discuss the processes undertaken by the ALFIE system and quantitative results of the final output from the system in terms of classification accuracy and network completeness. The authors conclude that the adoption of an object-oriented geospatial database has allowed complex discriminating characteristics of objects to be dynamically extracted, effectively enabling objects to classify themselves.